I read several posts here, saying RTK2 is not efficient or so, but this is not my purpose here. I would first like to **understand** RTK2 structure.
I start with chapter 1 and 2, giving my understanding and issues I found.

- chapter 1 : Kana and kanji : this chapter provides On OR Kun reading for some kanjis, which are related to hiragana or katakana. To be remembered by brute-force, or visual way. Fine.

- chapter 2 : pure groups
DEFINITION : this identifies groups of kanji having their UNIQUE on-yomi reading derived from one single common primitive elements, according to Heisig's explanation.
If definition above is correct, remembering only the unique reading of each primitive should be enough. As in the example, remembering that 中　is チュウ should be enough. Not even need to remember the list of kanjis which contain 中 (we already know them from RTK1), nor any sample compound.
ISSUE :
I found that my understanding is not true. For example frame #84 情　is said to be セイ、because in the pure group of 青。But frame 84 has a link to frame #1530 where same kanji 情 is has another on-yomi ジュウ。
So actually kanji containing 青 don't seem to create a pure group based on my understanding of the definition of "pure group"

It seems I missed something ? Or the above example is rare exception ?
Any help would be greatly appreciated...

You've misunderstood pure groups; they don't define unique on-yomi. What it means is that for the primitive, every kanji using that primitive will have *a* reading in common. It may not be the only reading for each kanji, though.

So any kanji with 青 will have *a* reading of セイ, but it may not be the only reading. (In fact most kanji with that primitive also have ジョウ as a reading, but not all.)

This is one of the major weaknesses of RTK 2, in my opinion. 情, for instance, has both セイ and ジョウ as on-yomi. But ジョウ is *much* more common; if you see a word with that kanji that you're not sure how to pronounce, you should always guess ジョウ rather than セイ.

yudantaiteki Wrote:You've misunderstood pure groups; they don't define unique on-yomi. What it means is that for the primitive, every kanji using that primitive will have *a* reading in common. It may not be the only reading for each kanji, though.

So any kanji with 青 will have *a* reading of セイ, but it may not be the only reading. (In fact most kanji with that primitive also have ジョウ as a reading, but not all.)

This is one of the major weaknesses of RTK 2, in my opinion. 情, for instance, has both セイ and ジョウ as on-yomi. But ジョウ is *much* more common; if you see a word with that kanji that you're not sure how to pronounce, you should always guess ジョウ rather than セイ.

Thank you!
And I understand the weak point you mention. Actually in my iPhone dictionary 情 the only on-yomi for it is ジョウ , even though the 風情　entry is included in the dictionary... but as an alternative reading ふう　じょう。。。
Then I understand also the purpose of the compound in this chapter...

yudantaiteki Wrote:You've misunderstood pure groups; they don't define unique on-yomi. What it means is that for the primitive, every kanji using that primitive will have *a* reading in common. It may not be the only reading for each kanji, though.

So any kanji with 青 will have *a* reading of セイ, but it may not be the only reading. (In fact most kanji with that primitive also have ジョウ as a reading, but not all.)

This is one of the major weaknesses of RTK 2, in my opinion. 情, for instance, has both セイ and ジョウ as on-yomi. But ジョウ is *much* more common; if you see a word with that kanji that you're not sure how to pronounce, you should always guess ジョウ rather than セイ.

It seems that the rule you mention does not even work : for example the pure group of白　（ハク） does not contain 的　(まと, テキ) nor 百　（もも、ヒャク)

So any kanji containing 白　does not have necessarily ハク as *a* on-yomi reading...

so my conclusion is that I need to remember ：
- the primitive and its onyomi
- the list of kanji or compound which are in each pure group

Contrary to Heisig sentence p20 about 中　pure group : "in other words, each time you see this primitive element you can be sure what the Chinese reading will be" is little bit confusing...

Or i am still missing something about pure groups ?

Yeah. In 的, the right-hand side (勺) originally served as the phonetic part, not 白. However even that doesn't work anymore in Japanese since 勺 (a rare kanji) has on'yomi シャク.

So no, it's not a rule you can rely on. Sometimes you even find multiple signals in the same kanji, like 恐 (工 and 凡), neither of which match the reading キョウ.

You can sort of get a intuition for when a part of a kanji is actually phonetic. Like in 的 the 白 is on the left and more cramped, like usually happens with signific parts (e.g. 人 in 住). Or with 恐, the top is actually a combined character 巩, which serves as the phonetic signal for キョウ.

So make of that what you will. I still see value in the RTK2 method, but you have to be very aware of its limitations. See the signals as just a loose hint to help boost learning, don't rely on them.

Personally I haven't used the RTK2 but I used a homebrew variant of the Movie Method and I assume the reading is X based on a certain primitive but then remember exceptions.

e.g. 召=ショウ, but when I see 超 I remember it's an exception and is チョウ instead. So I don't think that (all-召-except-超) are ショウ, but (all-召) are ショウ, and when I run across 超 I associate with an exception チョウ. That is, I may not be able to list all the exceptions for you, but I know an exception when I see one.

Other example: 燐 (right side, so 隣鱗, etc) = りん, but when I see 憐 I associate it as an exception having れん.

dmichaud Wrote:

yudantaiteki Wrote:You've misunderstood pure groups; they don't define unique on-yomi. What it means is that for the primitive, every kanji using that primitive will have *a* reading in common. It may not be the only reading for each kanji, though.

So any kanji with 青 will have *a* reading of セイ, but it may not be the only reading. (In fact most kanji with that primitive also have ジョウ as a reading, but not all.)

This is one of the major weaknesses of RTK 2, in my opinion. 情, for instance, has both セイ and ジョウ as on-yomi. But ジョウ is *much* more common; if you see a word with that kanji that you're not sure how to pronounce, you should always guess ジョウ rather than セイ.

It seems that the rule you mention does not even work : for example the pure group of白　（ハク） does not contain 的　(まと, テキ) nor 百　（もも、ヒャク)

So any kanji containing 白　does not have necessarily ハク as *a* on-yomi reading...

Doesn't he specifically say that 白 has to be on the right side of the kanji for it to qualify?

dmichaud Wrote:Contrary to Heisig sentence p20 about 中　pure group : "in other words, each time you see this primitive element you can be sure what the Chinese reading will be" is little bit confusing...

You're not alone if you found it confusing. The explanation isn't very clear. It seems to be a description of the 中 pure group only ("this primitive...the Chinese reading"). If so, it would have been better to explicitly point out that some of the kanji in other pure-groups have other readings (and direct the reader's attention to those little cross reference numbers).

In Ch 3 "One-time Chinese Readings", he does mention that while the kanji may have other readings, these particular readings are not shared by any other kanji. 白 （びゃく）　is in this list.

He mentions in the book's intro that the appearance and position of the primitive must be identical. I think an example would have been helpful as well as a reminder in main text.

As you're probably already aware, his categories are based on the list of RTK1 kanji, so you'll find exceptions among non-RTK1 kanji.

(P.S. I wouldn't bother trying to figure out the bizarre kun yomi chapter. Just skip it altogether.)

Thora Wrote:In Ch 3 "One-time Chinese Readings", he does mention that while the kanji may have other readings, these particular readings are not shared by any other kanji. 白 （びゃく）　is in this list.

He mentions in the book's intro that the appearance and position of the primitive must be identical. I think an example would have been helpful as well as a reminder in main text.

Yes page 6, "sixth, the use of kanji..." I found this sentence. But then i would have expected some hint before each pure group.
If we take again the example of haku 白　（ハク）, indeed the primitive seems to be on the right, and with nothing on top.
Thanks a lot for the hint to you and yudantaiteki !

Unfortunately Heisig in my edition does not mention anything about position of 白 for the pure group members...
But if you look one of the next groups 可　（カ）, the primitive is on the right （何、河）, at the bottom-right (荷）, on the left （歌）。。。

So it means that we must check and find-out, and then remember signal primitive + position -> reading. It looks like to do this, I will need some extensive cross table "pure group signal primitive" -> "RTK1 kanji"... has anyone already got this ?

Thora Wrote:As you're probably already aware, his categories are based on the list of RTK1 kanji, so you'll find exceptions among non-RTK1 kanji.

Yes this i understood, but thanks for the reminder !

mcaruso Wrote:Yeah. In 的, the right-hand side (勺) originally served as the phonetic part, not 白. However even that doesn't work anymore in Japanese since 勺 (a rare kanji) has on'yomi シャク.

So no, it's not a rule you can rely on. Sometimes you even find multiple signals in the same kanji, like 恐 (工 and 凡), neither of which match the reading キョウ.

You can sort of get a intuition for when a part of a kanji is actually phonetic. Like in 的 the 白 is on the left and more cramped, like usually happens with signific parts (e.g. 人 in 住). Or with 恐, the top is actually a combined character 巩, which serves as the phonetic signal for キョウ.

So make of that what you will. I still see value in the RTK2 method, but you have to be very aware of its limitations. See the signals as just a loose hint to help boost learning, don't rely on them.

Thanks. I try to come to the same ideas. I was expecting too much "rigor", or too much from the beginning, maybe. We can start by learning pure groups "as if all kanjis using the signal primitive" took this sound, and then add the exceptions on top of that later with other chapters. That's why learning sequence is important i believe...